x86_64: page fault with access violation error complaining supervisor thread not allowed to rwx

[chen-png]

To Reproduce
Steps to reproduce the behavior:

1. west build -b up_squared -p auto tests/kernel/device
2. west flash ~/up_squared.sh
3. see error

Logs and console output
Booting Zephyr OS build zephyr-v2.5.0-239-ge287fd7ab2fb
Running test suite device

START - test_dummy_device_pm
SKIP - test_dummy_device_pm

START - test_build_suspend_device_list
SKIP - test_build_suspend_device_list

START - test_dummy_device
E: Page fault at address 0x158 (error code 0x0)
PASS - test_dummy_device
E: Linear address not present in page tables

E: Access violation: supervisor thread not allowed to read
START - test_enable_and_disable_automatic_idle_pm
E: PTE: not present
E: Page fault at address (nil) (error code 0x2)
E: Linear address not present in page tables
E: Access violation: supervisor thread not allowed to write
E: PTE: not present
E: RAX: 0x0000000000000000 RBX: 0x000000000011a280 RCX: 0x0000000000000001 RDX: 0x00000000000000ff
E: RSI: 0x000000000011ada0 RDI: 0x000000000011a280 RBP: 0x00000000017c4ea0 RSP: 0x00000000017c4e90
E: R8: 0x0000000000000001 R9: 0x8000000000000006 R10: 0x0000000000105bb9 R11: 0x0000000000000000
E: R12: 0x000000000011a280 R13: 0x00000000017bd000 R14: 0x0000000000002000 R15: 0x0000000000105bb9
E: RSP: 0x00000000017c4e90 RFLAGS: 0x0000000000010086 CS: 0x0018 CR3: 0x0000000001dca000
E: RIP: 0x0000000000109676
E: >>> ZEPHYR FATAL ERROR 0: CPU exception on CPU 1
E: Current thread: 0x11ada0 (main)
E: Halting system

Environment (please complete the following information):

• Toolchain: zephyr-sdk-0.12.2
• Commit ID: e287fd7

[chen-png]

for testcases "kernel.common.tls", "kernel.message_queue", failed with the same error "E: Access violation: supervisor thread not allowed to read".

[dcpleung]

This is also reproducible on qemu_x86_64.

[dcpleung]

Tried with #32557 and seems to fix the issue. So wait for that to be merged...

[andyross]

Indeed, this is likely a dupe of the abort menagerie of troubles. I'm doing a ton of iterated qemu_x86_64 runs right now and not seeing this.